Statsmodels

Statsmodels is a Python package that provides classes and functions for the estimation of many different statistical models, as well as for conducting statistical tests and statistical data exploration. With over 10,300 stars as of 2026, Statsmodels complements SciPy and NumPy for users and developers of quantitative analysis in Python, offering an extensive list of result statistics and visual diagnostic tools. The library provides implementations for: linear regression models (OLS, WLS, GLS, recursive least squares, rolling regression), generalized linear models (GLM with Gaussian, Binomial/logistic, Poisson, Gamma, Inverse Gaussian, Negative Binomial, Tweedie families), generalized estimating equations (GEE) for longitudinal data, robust linear models with M-estimators (Huber, Andrew Wave, Hampel, RLM), linear mixed effects models (LME), generalized linear mixed effects models (GLMM), discrete choice models (Logit, Probit, MNLogit, Poisson, NegativeBinomial, Count models), time series analysis (ARIMA, SARIMAX, VAR, VECM, exponential smoothing: Holt, Holt-Winters, unobserved components, statespace models, Markov switching, GARCH), survival analysis (Kaplan-Meier, Cox proportional hazards), nonparametric methods (kernel density estimation, lowess), ANOVA, principal component analysis (PCA), factor analysis, mediation analysis, multiple imputation, power and sample size calculations, and over 80 statistical distributions (continuous, discrete, multivariate). Statsmodels integrates tightly with pandas DataFrames and patsy formula mini-language for R-like formula specification (e.g., 'y ~ x1 + x2'). Licensed under BSD-3-Clause.